When Meridian Labs decided to rebuild their enterprise onboarding flow, they didn't start with AI. They started with a question: why does it take our customers three weeks to see value?
The Problem
Meridian's platform is a data analytics tool used by enterprise teams. Their onboarding was a 14-step wizard that walked users through connecting data sources, configuring dashboards, and setting up alerts. Completion rates were dismal—only 23% of new users finished the flow.
The product team had iterated on this wizard for two years, optimizing copy, reducing steps, adding progress indicators. Nothing moved the needle significantly.
The Insight
VP of Product Sarah Chen made a crucial observation during a user research session: "Users don't want to set up a tool. They want to answer a question." The onboarding was designed around the product's architecture, not the user's intent.
The AI-Powered Approach
Instead of a wizard, the new onboarding asked one question: "What do you want to know about your data?" From there, an AI agent would guide the conversation, automatically connecting relevant data sources and generating an initial dashboard tailored to the user's stated goal.
Time-to-value dropped from 21 days to 8. Onboarding completion rose to 67%. But the most surprising metric was retention—users who went through the new flow were 3.2x more likely to become weekly active users.